correction for variables in a correlation

Hi everybody,

I am new here. I have a question I am unable to solve myself, but it seems so easy.
I have a dataset with slaughterdata from pigs. This dataset contains, amongst others, slaughterage (in days), slaugtherweight (in kg), backfat (in cm) and loin depth (= muscle thickeness, in cm).
As slaughterage and slaughterweight both have an effect on backfat and loin depth deposition, I want to correct for them. My endgoal is to have a correlationmatrix in which backfat and loin depth are compared to climate values.

How can I correct for age and weight to purely look at the effect of climate on backfat and loin depth?

Thanks in advance


TS Contributor
You can do a multiple linear regression of backfat (or loin depth, respectively) on age, weight, and climate.
You have adjusted for age and weight then. Although I do not know whether it makes sense to adjust
for total weight if you want to predict a weight component. Or are you maybe looking for something like
the ratio cm backfat per kg slaughterweight as dependent variable?

With kind regards



Less is more. Stay pure. Stay poor.
To follow-up, it also sounds like you have two dependent variables you would like to predict. Meaning you will need to run two models or "multivariate multiple regression". If you do the former, it may be good practice to make the alpha level a little smaller, since you are looking at two different outcomes and your by chance association risk increases in doing this.


No cake for spunky
You can also run MANOVA that deals with multiple dependent variables if you are brave and think that method is valid (it does not get used a lot as far as I know).